Introduction

PH345: Winter 2025

Phil Boonstra

Case Study 1

https://www.esri.com/arcgis-blog/products/product/mapping/mapping-coronavirus-responsibly/

  • Chloropeth: graduated colors

  • Provinces are different sizes, different populations

  • Comparison across provinces is difficult

  • Bar chart of number of cases for each Chinese province

  • Is a map justified?

  • Blue replaces red. Less “emotive”

  • Rates replace totals

  • Hubei province rightly set apart from others

  • Dots representing 10 cases randomly placed in each province

  • Potential misleading conclusion that Hubei province was overwhelmed

  • Totals represented by proportional circles

  • Not adjusted for population

  • All areas represented, e.g. Macau and Hong Kong

  • Log-transformed totals

  • Importance of legend

  • Logarithm de-emphasizes extremely large values but risks over-emphasizing small values

  • Inappropriate ‘smoothing’ of data based upon geographic center

  • Epicenter (Hubei) is lost

  • Suggests all of eastern China was overwhelmed

  • Choice of projection

  • Web Mercator: up is always north. Distortions lead to risk of misinterpreting geographic area

  • Albers Equal Area preserves geographic area but can distort shape

Case Study 2

https://www.worldbank.org/en/publication/globalfindex/Report

  • Light blue = 2011,2014,or 2017; Dark Blue = 2021

  • How difficult is it to…

    • …find specific country?
    • …find greatest change?
    • …summarize overall changes?
    • …precisely report a number?

Sixty-two percent of the unbanked cited ‘lack of money’ as one of multiple responses (figure 1.2.3). (p35)

  • Plots of proportions should clearly communicate what denominators and numerators are

In Bangladesh, 69 percent of unbanked adults have a mobile phone; in Nepal, 73 percent (figure 1.2.6).

Figure 2.4.2 shows the relationship between payment inflows and the use of financial services among payment recipients as developing economy averages.

Case study 3: Dr. John Snow’s map

(Gilbert, 1958)

https://www.jstor.org/stable/pdf/1790244.pdf

Course Objectives

  1. To understand the principles of effective and accurate graphical representation of different data types;

  2. To draw conclusions from graphical representations about relationships and trends in variables;

  3. To understand how graphical representations of data can be used to mislead or exaggerate relationships;

  4. To create and improve data visualizations using the R statistical environment;

References

Gilbert, E.W., 1958. Pioneer maps of health and disease in England. The Geographical Journal, 124(2), pp.172-183.